My oldest is entering kindergarten next year, and I wanted to answer two big questions:

What follows in this article is my dive into the deep end of school performance metrics, and I used them to make, what I hope is, an informed decision.

Good primer with general info: http://www.ncpublicschools.org/docs/accountability/reporting/2018/documentation/chrtgrdrt18.pdf

Two main testing metrics

If you look at the report cards prepared for Wake County schools by the State, you’ll see that they each have two metrics.

I was familiar with the EOG tests. I took them as a kid, and it’s a pretty straight-forward measure to interpret. An “A” means that kids got a lot of questions right and an “F” means they didn’t.

The growth score was new to me, but given how prominent the color of the school was compared to it’s letter grade, it seemed important.

Should I consider growth scores?

When I started, student growth seemed the most important factor. Consider two third grade teachers:

  • Teacher A:
    • Starts off with a class of children reading at a first grade level.
    • At year end, they are reading at a third grade level.
  • Teacher B:
    • Starts off with a class of children reading at a fourth grade level.
    • At year end, they are still reading at a fourth grade level.

Based on the EOG exam, Teacher B would look better because her children can read at a fourth grade level. Looking at student growth should give better insights into teacher performance.

The goal of EVAAS and other value-added modeling is to isolate the role of the teacher in a student’s growth and control for things like the student’s socio-economic status or previous competency level.

My first scatter plot of schools compared their EOG and growth scores. Along the x-axis is the percent of students that scored above level four on the EOG. Level 4 is “College and Career Ready” (CCR). The y-axis shows the growth score. Scores between -2 and 2 mean that students were learning at the rate that was expected. Schools above 2 exceeded this growth while schools below 2 saw students growing slower than expected.

There is a relationship between the two test scores (represented by the dashed blue line), but the correlation was slight. For the most part, the schools with the lowest EOG scores failed to meet their growth targets, and the schools with the best EOG scores met or exceeded them. With that said, the schools are scattered around widely.

After digging through several locations online, I was unable to get more than a general impression of how the growth targets were forecasted for a student. I was unable to find a statistical model published by SAS (creators of EVAAS). I reached out to a teacher friend of mine and quickly learned what teachers thought of EVAAS. To put it politely: they don’t like it. One of the biggest reasons behind this was a lack of transparency of how the scores were generated. Apparently, I’m not the only one who couldn’t figure it out.

Undocumented statistical models raise several flags. Perhaps there is validation work I’ve not seeing demonstrating that the models can be trusted, but I decided to put more weight on the EOG scores and use the EVAAS system as another point of reference (e.g. to break ties).

ITBS

One thing I learned quickly was that the EOG and EVAAS scoring didn’t apply to private schools, which complicated the comparisons. The private schools I contacted all used the Iowa Test of Basic Skills (ITBS), which at least facilitated comparisons between private schools. I’ll use these below.

Digging Deeper

The scores above are for the school as a whole. The socio-economic mix of schools is often a large factor in these overall scores. This means that understanding the makeup of schools. As in our example when considering growth scores, an excellent school may be in a disadvantaged area and look worse based on EOGs alone. The map below shows a breakdown of every public, public charter, and private school in Wake County. Click on a pie chart to see the stats as well as the school type.

Luckily, state scores are available for different demographics (race, gifted, etc.). This facilitates making fair comparisons.

Local schools

This criteria is mainly personal preference, but I wanted my son’s school to be near our home. This makes it a lot easier for me to attend functions, but also makes the every day pickup and drop off routine easier. If he takes the bus to school when he’s older, he won’t spend as much time on it as I did when I was a kid. I hated the bus.

With my new understanding of EOGs, EVAAS and ITBS, I decided to compare some local schools of various flavors.

First, it seemed like a good idea to see where the public schools fell on the scatter chart from above.

Sticking with my decision above to favor EOG scores over EVAAS, Penny Road appears to do slightly better than Oak Grove, but the charter school scores better than either (and on both metrics).

The chart below shows the EOG scores by school and race. Black and hispanic students get the best scores at Peak while white students score higher at Penny Road.

Reproducibility and Data

Other Resources

I encountered many great resources that helped inform my investigation. Every public school (non-charter) in Wake County has a School Improvement Plan (SIP) that you can view online. This is a candid, introspective look by the school at what they need to do to improve. Each link below provides the login information to view that school’s SIP on a system called Indistar.

Oak Grove: https://www.wcpss.net/Domain/6942

Penny Road: https://www.wcpss.net/Domain/7040

Another great resource is https://www.greatschools.org/. You can get detailed information for each school along with reviews. For example: https://www.greatschools.org/north-carolina/cary/1975-Penny-Road-Elementary/.